On Fri, 14 Feb 2014, gareth wrote:

Scouting around (plagiarising?) published designs for HR RXs and sundry
boatanchors, it seems that where there are tuned RF stages, then 2 tank
circuits seems to be the norm, and then only 3 ccts if the RX IF
is ooo 455kHz when the coverage needs to extend to 30MHz.

Looking at the rate of progress here (or lack of it!!) I am trying to
conceive of a scheme to use up all the tuning condensers (capacitors
only came in AFTER some of them were manufactured :-) ) in
the junk box in one final big push to releve the executors of my will
of difficult decisions :-)

One reason for the superheterodyne receiver was to get away from the TRF
reciver, which was some stages of tuned amplification on the signal
frequency. Lots of problems once you started getting enough
amplification.

With the superhet, the RF stage before the mixer is there to offset the
noisy mixer (which for most of the HF range isn't a problem) and to get
rid of any images.

The cheap ham and shortwave receivers of the old days would tend to have
only one RF stage, and you'd see reviews for the commercial receivers with
things like "the image was as strong as the desired signal". Some
military receivers improved on that, and the HRO receiver was well known
for better image rejection at the higher frequencies despite having a
455KHz IF. It had not only two stages of RF amplification, but apparently
well designed tuned circuits.

Double conversion was a better way to get rid of images. The first
double conversion receivers tended to be to a low HF frequency, say about
2MHz, with a tuned circuit or two, then a drop to 455KHz for selectivity.
That wave of receiver generally didn't have a lot of selectivity at the
first IF frequency. Indeed, in a number of receivers, the mixer stage
became another 455KHz IF amplifier on the lower frequencies where the
double conversion didn't kick in. (Or to look at it in reverse, one of
the 455KHz IF stages became a mixer on the higher frequencies when double
conversion kicked in.)

The other method at the time, if you had the money, was like the Collins
75A series receivers, which in effect were a tuneable receiver converting
to 455KHz, with a crystal controlled converter ahead of that tuneable
receiver. Those were all the same basic scheme, though some tuned a ham
band directly, while others (like the Collins) tuned some other segment
with the converter needed for all ham bands).

But both double conversion schemes were as good as the tuned circuits.

It was only when HF crystal filters came along that image rejection became
much simpler. Pick a 9MHz filter, and the image is 18MHz away, not much
needed to reject the image. At that point, you could do away with the RF
stage, or at least keep it to a minimum, maybe just enough gain to offset
the front end circuits.

It's better to kick in amplification when needed than keep it in the
circuit all the time, which is how more recent designs have gone.

Once filters above the shortwave spectrum were viable, that really put the
image frequency far away from the signal frequency. Some of the early up
converting receivers just used a low pass filter ahead of the mixer (or
had that as one of the options, I think the Racal receiver with the Wadley
loop had this option, but I may be misremembering).

With such a high IF, it was simpler to get rid of images.

INdeed, one could basically build a receiver and leave off the front end
selectivity. Build the receiver well to that point, then play with what
comes between the antenna and the mixer. Got a really tough RF
environment? Maybe a crystal filter at the signal frequency is the
answer, so long as you only need a very small band of the spectrum. Put in
simpler filtering for the bands that you don't care about, make it better
for the one you do care most about. Or live with a low pass filter for
general coverage. For a band that you might care about most, have a good
filter that can be peaked on the signal, for some other band (or where you
need to tune for a converter for the VHF/UHF bands) you might want a
broadband filter that covers the 500KHz of the band, not requiring any
tuning within that band.

The point is that once you make the image rejection easier, you aren't
fussing with image rejection, you are working on other things that is more
important.

Michael